Ahmad Fattahi

How much data science and advanced analytics is being done in the PI community?

Blog Post created by Ahmad Fattahi Employee on Apr 2, 2018



My colleague, Brian Bostwick, and I have been asking ourselves this question: how much data science and advanced analytics is being done in the PI community? As a measure we decided to look at the activities at the upcoming PI World in San Francisco. And the result was overwhelming! So much so that we started crafting a list to build a mental map of what's happening between the PI System on one side and Data Science and Advanced Analytics on the other. It turns out the manufacturing world is very much up to speed already. What we found was that leveraging data science in manufacturing is quite ubiquitous and wide-spread across multiple industries such as Oil and Gas, Power and Utilities, Forest and Paper, Industrial IT, Life Sciences, Water, and Metals and Mining.


Another observation is that there is quite a bit of low hanging fruit; i.e., implementing simpler methods, such as linear regression or decision trees, can yield huge improvements to the bottom line. Simpler and better explained algorithms carry other benefits such as easier reproducibility of the process. We also learned that many users do feature-engineering within the PI System right where the time series data lives. This architecture gives them an edge because they can scale the process nicely, avoid moving data unnecessarily or creating new silos.


Anyway, we got so encouraged that we decided to share the outcome with the whole community. The list includes offerings by customers, partners, and OSIsoft employees. You can find great industrial use cases (earlier days in the week) as well as learning opportunities in the form of talks and hands-on labs (later days in the week). Hope it can help you navigate the plethora of activities at PI World. More details can be found on the official event website. Please note that in case you notice a discrepancy between the agenda here and the public event website the latter should be taken as the final authority.


Day 1 - Tuesday, Apr 24:


Day 2 - Wednesday, Apr 25:


Day 3 - Thursday, Apr 26:


  • Labs
    • 10:30-12:30 - Hilton: Learn How to Add Automated Machine Learning to the PI System Quickly and Easily (Sponsored by Falkonry)
    • 1:30-4:30 - Hilton: Apply Predictive Machine Learning Models to Operations
    • 1:30-4:30 - Hilton: A Digital Plant Template for operational Insights - An Enterprise Strategy
    • 1:30-4:30 - Parc 55: Advanced Analytics for PI data for Data Scientists


Day 4 - Friday, Apr 27:

  • Labs
    • 9:00-12:00 - Hilton: Fit for Purpose - Layers of Analytics using the PI System – AF, Matlab, Machine Learning
    • 9:00-12:00 - Parc 55: Introduction to Data Science for PI Data for PI Professionals